PAKISTANINSTITUTEOFDEVELOPMENTECONOMICS

EstimatingtheMiddleClassinPakistan

DurreNayab*

October,2011

*TheauthorwouldliketoexpresshergratitudetoDr.RashidAmjad,ViceChancellor,PIDE,andMr.HasanRabayfortheircommentson anearlierdraftofthispaper.TheauthorisalsothankfultoDr.NadeemulHaque,formerViceChancellor,PIDE,formotivatingmetotake upthistopic.Usualdisclaimerapplies.

Estimating the Middle Class in Pakistan

The concept middle class is one of the most commonly used terms in the social sciences, including economics, sociology and political science. Despite its frequent use there is, however, no consensus on what the term exactly implies and its meaning remains ambiguous depending primarily on the context in which it is used. It is viewed as the class that is between, and separates, the lower and the upper classes, that is the rich and the poor, but there is no agreement on the exact boundaries that separates them. Most of the definitions and measurements of the middle class continue to be somewhat arbitrary and vague.

Historically, the concept of class has been there for long but the very recent interest in the middle class has stemmed from the emergence of this class in the Asian economies, especially India and China, which continued to grow even during the global recession. Consumer spending in the developing Asia continued to grow during a period when the developed world was struggling to keep their economies afloat at both the national and household levels. It was an important development, as the middle class has always been deemed critical to a countrys socio-economic and political growth and ensuing stability. Hence, not surprisingly the middle class has held a centre stage in most economic discourses, and depending on the stage of its development and state of the economy the middle class has been attributed to be, among other things, growing stressed, shrinking, powerful, threatened, burgeoning, mobilised, rising, or marginalised. In a world of globalised economies this raises the crucial question, who constitutes the middle class?. An Indian schoolteacher with an annual income of $2,500 is considered middle class but for an American family the amount may have to be around $200,000 before it considers itself middle class (Aho, 2009) making any income based universal definition meaningless.

The concept upper, middle and lower class was mainly developed in sociology in reaction to the more rigid and deterministic Marxist concept of class but the definition remains ambiguous. For the estimation of middle class, however, it is necessary to first be clear about what exactly constitutes being a middle class. The present paper first briefly describes the different ways of analysing and defining class, followed by exploring the importance of the middle class and the 1

various means of estimating its size. The paper then suggests a refined measure of estimating the middle class in Pakistan, a measure that captures all possible aspects of the class given in its various definitions. Before reaching conclusions, the paper also explores the regional and occupational context of the middle class. Concept of Class and the Notion of the Middle Class As very aptly put by Streans (1979: 377), The concept middle class is one of the most enigmatic yet frequent in the social sciences. Part of this difficulty can be ascribed to the ambiguity regarding the very concept of class. Classical perspective on the concept of class can be traced back to Karl Marx with his classification based on the relationship to the means of production and Max Webers categorisation on the basis of wealth, prestige and power1. In both Weberian and Marxist traditions there is a rejection of any simplistic gradational definition of class as they link peoples social relations to economic resources of various kinds. These schools view as social relations having an impact on the material interests of people and forming the foundation of conflict and cohesion. Marx and Weber, however, basically differ in their theoretical orientation, reflected in Marxian notion of exploitation and Weberian life-chances.

There is a vast body of literature on the concept of class having its roots both in the Marxian and Weberian schools of thought with each arguing about inclusion of certain properties in defining classes. For instance, classes have been defined by: position and role (Mosca, 1939; Pareto, 1963); authority relations (Dahrendorf, 1959); status rank (Warner, 1960;Lenski, 1966); inter-marriage (Schumpeter 1951); cross-classification of property and authority (Ossowski, 1963; Wright, 1979); degree of structuration2 (Giddens, 1973); and property, employment and authority relations (Goldthorpe 1987, Erikson and Goldthorpe 1992). It would not be wrong to state that all these definitions of class in a broad way follow the Marxian and Weberian concept of class pursuing their respective notion of exploitation and life-chances. That is, intra-class conflict

ForausefulsummaryonbothschoolsofthoughtreadWright(2003). Giddens proposed that what humans do is dependent on the preexisting social structures that are governed by norms. All human actions are restricted by the elements that create the social structures. Giddens believed that structures (traditions, institutions and expectations) are universally steady but could be changed mainly by some unintentional consequences of action, for instance, when people begin to pay no attention to the social norms, replacethem,orfollowtheminadifferentway.
2 11

leading to exploitation of one by the other, and class as a source of certain qualities and actions ensuing in the associated life-chances. Wright (2003, pp: 1-3) provides a useful summary of the different underlying themes resulting in the variety of class concepts. He divides these themes in five categories which are as follows: 1. Subjective position-the way people locate themselves and others in a social structure characterised by inequalities. 2. Objective position-distribution of people objectively on a gradational scale of economic inequality, represented mainly by income or wealth. This results in classes such as upper class, middle class, upper middle class, lower class and underclass. 3. Relational explanation of economic life chance-also characterising the Marxian and Weberian schools of thought, it explains inequalities in economic life-chances including standards of living. Based on the relationship of people to income-generating resources or assets, class becomes a relational, rather than a gradational concept. 4. Historical variation in systems of inequality- focussing on the macro-level instead of microlevel, it explains the variation across history in the social organisation of inequalities. 5. Foundation of economic oppression and exploitation-looks into the nature of changes needed to eliminate economic exploitation within societies.

So where does in the above-mentioned scheme of things fall the middle class? In relating it to the means of production, as done by Marx, we can consider the middle class to be falling somewhere in between those who own the means of production (the bourgeoisie) and those who must sell their labour for survival (the proletariat). A class that came to be referred to as the petty bourgeoisie, includes small producers/proprietors, like shopkeepers and small manufacturers. Similarly, in the Weberian notion of class, middle class can occupy the mid-position on the continuum of wealth, power and prestige. In the wealth continuum middle class can be presented by individuals who are neither rentiers nor unskilled labourers. On the power continuum they can be the people who are not as weak as to carry out the commands of others but not as influential to achieve their goals despite opposition. Similarly, they cannot be individuals who receive little respectful treatment nor the ones who are entitled to deferential

and respectful treatment. The middle class is, thus, on the middle rung of all Weberian continuums.

Sridharan (2004), while analysing the Indian middle class, believes that the classical approach to the middle class is inadequate for analysing contemporary societies, especially those in the developing world. He argues that for over a century the growth of economies, specifically capitalist ones, has relied on advancements in knowledge and improvements in the regulatory state machinery leading to a growth of white-collar occupations, salaried and self-employed, marked by increasing degree of education. The income of this new group of professionals normally falls somewhere between those of the capitalists (the equivalent to Marxian bourgeoisie) and the manual workers(the Marxian proletariat), and signify the new middle class. The size and influence of this class has increased with time, affecting the political and ideological aspects of society. Their significance also grows as economies move towards the services sector.

This new middle class, however, is not considered a real class by Wright (1985). He considers it a contradictory location within class relations (p: 9). Wright considers this new class of professionals and white-collar workers different from the petty bourgeoisie comprised of small independent producers and shopkeepers. Relative autonomy of this new middle class, especially in the developing countries, distinguishes it from other classes- an autonomy that originates from the possession of skills and say in public affairs. Bradhan (1989, 1994) shares this view and considers professionals, including all white-collar workers, as the dominant proprietary class because of the level of human capital manifested in their higher education, technical expertise and possession of scarce skills. The dominant proprietary classes, thus, now comprise industrial capitalists, big farmers and the professionals, both civil and military. What makes Middle Class Important The middle class is increasingly considered a precondition of stability in the social structures, a means of mitigating inequalities in a society, and a pathway to growth and development. This idea has gained strength from the events in China and India where the burgeoning middle class is believed to be holding the future of these countries. It may be mentioned here that the

importance attributed to the role of the middle class is not a recent phenomenon. For instance, Landes (1989) talks about Englands early dominance in terms of the English middle class of the 18th and 19th centuries. More recently, Birdsall, Graham and Pettinato (2000) consider middle class the backbone of both market economy and democracy in the face of globalisation. Likewise, Easterly (2001) after analysing a large number of countries concluded that nations with a large middle class tend to grow faster, at least in situations of ethnic homogeneity. "Thus it is manifest that the best political community is formed by citizens of the middle class, and that those states are likely to be well-administered in which the middleclass is large .... where the middleclass is large, there are least likely to be factions and dissension." Aristotle 306 BC (quoted in Easterly, 2001:1) The above-stated stabilising role of the middle class originates from the buffer role it seems to play between the polar tendencies of the lower and upper classes. Easterly (2001), for instance, shows that a higher share of income for the middle class is linked with higher growth, more education, better health and less political instability and poverty in a society. These qualities make a decline in the middle class a potential threat to economic growth and political stability. Esteban and Ray (1999), for example, show an occurrence of more frequent societal conflict in the presence of a weaker middle class.

The middle class is increasingly seen as a group gaining political influence that can be associated with the progressively larger role they are playing in the public and services sectors. The middle class is also linked with the nature of government a country has, as shown by Moore (1966) in his classical work associating democracy with the middle class. This idea is supported by Collier (1999) when he studies various democracies finding the middle class to have allied with the lower classes to push for an inclusive political system. On the contrary, however, he also found instances where the middle class formed an alliance with the upper class putting up with a restricted democracy or even a dictatorship. A similar relationship was also found by Leventouglu (2003) when he observed an ambivalent behaviour of the middle class during political transition. Depending on the situation, the middle class could act as an agent of change or work for maintaining the status quo. If the middle class believed that their children would retain their middle class status then they would not resent semi-democracy or 5

even encourage an autocracy to block any redistribution. On the other hand, if the middle class is not guaranteed their status they would strengthen the lower class so as to push for redistribution under democracy. This ambivalent behaviour makes the role of the middle class even more politically important, and as pointed out by Acemoglu and Robinson (2003, p: 8-1) the decisive voters in democracy are often from the middle class.

All other roles of the middle class granted, including those discussed above, the most significant is the one that links it with the growth and development of economy. Bannerjee and Duflo (2007) provide a useful summary of the relationship between the middle class and economy. Theorising back to Weber and using a vast body of literature, they delineate three reasons for considering the middle class vital for economies. These include: 1. New entrepreneurs emerge from the middle class who create employment and opportunity of growth for the rest of society. 2. The middle class with its strong values stresses on the accumulation of human capital and savings. 3. The middle class consumer is willing to pay a little extra for quality, thus, encouraging investment in better quality production and competitive marketing, which spurs higher level of production and leads to increasing income for everyone.

All these aforementioned factors make the middle class vital for any economy. There are, however, words of caution. Nirvikar (2005) and Basu (2003) while commenting on the middle class consumption pattern warn that although consumer spending enhances aggregate demand and stimulates the economy in the short run, it does not necessarily translate into higher sustainable growth. They also voice concern about the sustainability of these high levels of consumption, and the depressing effect they have on savings, and hence consequentially on investment. Measuring the Middle Class After studying the middle class in various countries, Stearns concluded, Recent theory does not even include a satisfactory statement on the size of the middle class, caught as it is between the dazzling power above and massive numbers below (1979: 378). The desire to statistically

measure socio-economic phenomenon, like the middle class, believes Cole (1950), originates from the success of such exercises in natural sciences. The task of quantification, however, is much tougher in social sciences due to the complexity and variability of social realities. Unlike facts in natural sciences, with their universal applicability, social realities dealt by social sciences cannot be generalised even at the micro level, leave alone universally. Despite this handicap extensive literature can be found studying the middle class not just theoretically but empirically as well.

For the identification and measurement of the middle class, mainly two methods are used to define who is included, and who is not, in the middle class. One way to do it is in relative terms and the other one is the absolute way. The second issue is the premise of inclusion, be it in relative or absolute terms. The most commonly used dimension is purely economic in nature as it bases the rationale of inclusion, or exclusion, in the middle class solely on personal income or expenditure. There is however criticism on this over reliance on income and for ignoring other factors, like occupation, wealth and labour market relations (Goldthorpe, 2009).

If we look at the various ways of quantifying the middle class, as presented in Table 1, we find that most of the definitions are absolute in nature using purely economic rationale for inclusion in the middle class. The definitions use intervals of income values, with lower and upper limits, to measure the middle class or define it by using different median values of income. It would not be wrong to re-emphasise here the arbitrary nature of, and the lack of consensus on, the different ways of measuring the middle class, evident in the absence of agreement on the upper and the lower limits demarcating the boundaries for the middle class in the total population.

Table 1: Selected Definitions and Methods of Measuring the Middle Class

No. Author 1. Thurow(1987);Birdsall,Graham andPettinato(2000) 2. 3. 4. 5. 6. 7. 8. 9. Easterly(2001) MilanovicandYitzaki(2001) Ravallion(2009) Birdsall(2010) AcsandLoprest(2005) Wheary(2005) Nehru(2010),Yuan,etal(2011) BanerjeeandDuflo(2007) Approach Relative Relative Absolute Absolute Absolute Absolute Absolute Absolute Absolute Relative Relative Absolute Absolute Relative Relative/ absolute Rationale Economic Economic Economic Economic Economic Economic Economic Economic Economic Economic Economic Economic Economic Definition 75%to125%ofthemedianincome Expenditure quintilestwotofour IncomeofPPP$12$50/day/person ExpenditureofPPP$2$13/person/day Incomeof PPP $10/person/daybutnot inthetop5% DoublethePovertyLineofPPP $2/person/day Doublethenationalpovertyline IncomefromPPP$2to$20/person/day ExpenditureofPPP$2$10/person/day Doublethe medianincome Threetimesthemedianincome AnnualhouseholdincomebetweenPPP $13,500to$113,000 Expenditurefrom$2to$20/person/day

10. Peichl,ShaeferandSchneider (2008) 11. Brezenski(2010) 12. McKinsy(2010) 13. ADB(2010);Bhandari(2010); Chun(2010) 14. Gilbert(2003)1 15. Goldthorpe(1992)

Economic& Gradationbasedonincomeandnature Occupation ofoccupation Occupation Threemainclusterstheserviceclass, theintermediateclass,andtheworking class.

Source:FoundinReferences. Note:1.PPPPurchasingPowerParity 2.SeeFigureA1intheAnnexurefordetails.

Estimating the Middle Class in Pakistan This brings us to the question, How big is the middle class in Pakistan?.Using the Pakistan Social and Living Measurement Survey (PSLM), conducted in 2007-08, the paper measures the magnitude of the middle class through the definitions given above in Table 1. It may be mentioned here that it is a household, and not an individual, that by all these definitions is categorised as the middle class. Depending on the definition applied, it is found that the size of the middle class ranges drastically in the country, as can be seen from Table 2. Applying the definitions having solely an economic rationale, we find the middle class to range from 60 per cent of the population (Table 2, Definition One) to being totally non-existent (Table 2, Definition Five). Translating it in number of people, using the population base of 187 million as it stands on mid-year 2011 (USCB, 2011 and UN, 2009), the size of the middle class ranges from a huge 112 million to no one. This variability, as stressed earlier, reflects the complexities and arbitrariness associated with defining and measuring the middle class. Table 2: Estimation of the Middle Class in Pakistan by Various Existing Definitions
Class Lower Middle Upper Lower Middle Upper Lower Middle Upper Lower Middle Upper Lower Middle Upper SizeofClassesbyDefinitionUsed(%) Proportion(%) MiddleClassinNumber(inmillions)1 DefinitionOne:75%125%oftheMedianincome 32.8 29.5 55.2 37.7 DefinitionTwo:Quintileapproach 20.0 60.0 112.2 20.0 DefinitionThree:IncomeofPPP$12 $50perpersonperday 98.1 1.8 3.4 0.1 DefinitionFour:ExpenditureofPPP$2 $13perpersonperday 56.1 42.9 80.2 1.0 DefinitionFive:IncomeofPPP$10perpersonperdaybutnotintheTop5% 97.3 0.0 0.0 2.7
Continued:

Table2continued: Class Lower Middle+Upper Lower Middle+Upper Lowerlower(<$1.25) Lower(<$1.25$2) LowerMiddle($2$4) MiddleMiddle($4$10) UpperMiddle($10$20) Upper(>$20) Lower Middle Upper Below Above Below Above Lower Middle Upper Lowerlower(<$1.25) Lower($1.25$2) LowerMiddle($2$4) MiddleMiddle($4$10) UpperMiddle($10$20) Upper(>$20) SizeofClassesbyDefinitionUsed(%) Proportion(%) MiddleClassinNumber(inmillions)1 DefinitionSix:DoublethepovertylineofPPP$2perpersonperday 87.3 23.7 12.7 DefinitionSeven:Doublethenationalpovertyline 80.7 36.1 19.3 DefinitionEight:IncomefromPPP$2to$20perpersonperday 30.1 28.2 28.5 76.7 10.5 2.0 0.8 DefinitionNine:ExpenditureofPPP$2 $10perpersonperday 56.1 42.2 78.9 1.7 DefinitionTen:Doublethemedianincome 83.5 30.9 16.5 DefinitionEleven:Threetimesthemedianincome 92.1 14.8 7.9 DefinitionTwelve:AnnualHHincomebetweenPPP$13,500to113,000 93.9 6.1 11.4 0.1 DefinitionThirteen:ExpenditurefromPPP$2to$20perpersonperday 21.3 34.8 32.4 81.3 9.8 1.3 0.4

Source:CalculationsbasedonPSLM20072008 Note: 1. Estimation of the approximate size in numbers is based on using the midyear population of 187 million in the year 2011. Following these definitions, the size of the middle class in Pakistan has shown a growing trend (see A9) so applying the 20072008estimatesto2011populationcangiveanunderestimation/minimumsizebutnotanoverestimation. 2.UsingPPP(PurchasingPowerParity)ofUS$1=PakRs.24.47in2008.

Among all the definitions given above, Definition Eight and Definition Thirteen, based on gradation of income and expenditure per person per day, respectively, are currently the most extensively used measure employed to estimate the middle class (as also used by Chun (2010)

10

and Bhandari (2010) among others)3. This definition too, however, suffers from the same drawback of relying solely on one criterion. As also pointed out by Eisenhauer (2008), Atkinson and Bourguignon (1982), Kolm (1977), Bourguignon and Chakravarty (2003) and Gilbert (2003), being a part of the middle class should be ascertained by a persons socio-economic attributes holistically. Income is an important aspect but other qualities like level of health, wealth, education and specialised knowledge are also significant factors for constituting a class.

Technically speaking too, most of the definitions suffer from serious drawbacks. For instance, the quintile approach can be useful in measuring or comparing income or expenditure growth but cannot be used as a method to estimate the middle class as the size cannot shrink or expand and by definition would permenantly remain at 60 percent. Any denomination of the median income should also be used with caution in low income countries like Pakistan. Taking 75 per cent of the median income might lead to the inclusion of people below the poverty line in countries with very low income levels.

In the above-stated definitions and resulting estimates there are issues with the lower bounds set for inclusion in the middle class. While some of the definitions (like Definition Three and Five) set the limit too high4, resulting in a very small middle class or in the absence of a middle class altogether, there are other definitions that set the limit too low, like those that set the lower bound at $2 per person per day. Does the middle class begin where poverty ends? Ravallion (2010: 446) supports, the premise that middle class living standards begin when poverty ends. This paper, however, supports the argument forwarded by Horrigan and Haugen (1988:5) when they posit, to ensure that the lower endpoint of the middle class represents an income significantly above the poverty line. The middle class should, hence, include only those households that do not face the risk of experiencing poverty at all, and are not just those who are outside the the realm of poverty at a particular time. As also pointed out by Tilkidjiev (1998), it is not sufficient to be wealthy to be in the middle class, this paper also premises that middle income should not be considered middle class.
3

These definitions have been used to estimate the muchquoted Indian middle class to be as big as 250 million. Pakistansaround80millioncompareswellwithitsneighboursmiddleclass. 4 The much talked about Indian middle class also disappears following this definition, indicating the futility of applyingittothedevelopingeconomies.

11

The middle class has a multidimensionality attached to it and any useful measure should attempt to capture it. The middle class has certain intellectual, political and social connotations, along with economic ones, that differentiate it from the middle-income. While middle-income is purely an economic term, the middle class falls more in the sociological domain. The concept of class has many dimensions, including: economic, like wealth, income and occupation; political, including status and power; and cultural, such as values, beliefs, lifestyle, and education. Taking all these factors in consideration, a refined measure of estimating the middle class is proposed using a weighted composite index. The suggested measure is a combination of relative and absolute approaches applied to estimate the middle class, and the rationale is not solely economic but a collection of various socio-economic factors. After the formation of the subindices, the Principle Component Analysis (PCA) was carried out to calculate the weights given to each component of the composite index. The calculation of these sub-indices followed an absolute approach, while the final gradation of the resulting composite index into different classes was done with a relative approach. The suggested components of this composite index, their rationale and the method used to measure them areas follows: 1. Educational Index: In todays world, college education is the minimum for any individual to succeed (Wheary, 2010). With its low literacy levels, presence of even one person with college education (completed or currently in college) in a household qualifies it to be categorised as middle class. Quantification of the index: at least one individual in the household with college education= 1, no one in the house has college education = 0. 2. Income Index:5 Instead of using any upper and lower bounds or any percentage of median income, due to the shortcomings stated above, the double the poverty line

Income Index was created instead of an Asset Index, which by some is considered a better indicator of security against vulnerability (Sorenson, 2000; Brandolini, Magri and Smeeding, 2010; and Bradhan, 1989), because of the nature of the topic understudy and inclusion of some of the asset variables in other indices. For instance: ownership of house was part of the Housing Index; possession of movabledurable assets was included in the LifestyleIndex;andnonmaterialassets,likehighereducation,wasapartoftheEducationIndex.

12

definition is used to form the Income Index. Being twice as much away from the poverty line, these households are believed to be facing minimised risk of poverty. National poverty line was used for this purpose by inflating the officially accepted available poverty line to the year 2007-2008, which came up to Rs. 1084.20 per person per month. Quantification of the index: households double the poverty line= 1, households below double the poverty line= 0. 3. Housing Index: Possession of a house is vital for the middle class status (Brandolini 2010, Banerjee and Duflo 2007, Wheary 2010). Considering most of the houses in Pakistan are self-owned, in order to differentiate between the classes the housing index was disaggregated into ownership, persons per room, and availability of water, gas and electricity in the house. Quantification of the index: (i) ownership of the house = 0.3, otherwise = 0 (ii) number of persons per room- 1= 0.4, 2= 0.3, 3= 0.2, 4= 0.1, >4= 0 (iii) availability of electricity= 0.1, otherwise= 0 (iv) availability of tapped water in the house= 0.1, otherwise= 0 (v) availability of gas= 0.1, otherwise= 0. All the sub-components adding to a maximum of 1 and a minimum of 0. 4. Lifestyle Index: the middle class is associated with a certain lifestyle associated with expenditure on consumer durables- one of the primary reasons they are considered a boon for economy. These also form part of the movable assets possessed by the household. The PSLM asks the households about the ownership of twenty-three consumer items and this index includes all of them. Quantification of the index: each item owned= .0435. All the sub-components adding to a maximum of 1 and a minimum of 0. 5. Occupation Index: after income, occupation is considered as the most important factor affecting any individuals or households class categorisation. Occupations were divided in two categories in this index, namely, manual occupations and non-manual

13

occupations. A drift away from manual occupations is deemed imperative to be in the middle class in a vast body of literature, including Brandolini (2010), Gigliarano and Mosler (2009), Beteille (2001), Grant (1983) and Cole (1950). Quantification of the index: if the occupation of the head of the household or the person earning the most in the household is non-manual= 1, otherwise= 0. The sub-indices, comprising the composite index, were then weighted through the PCA method and their scores were added up to give the total score for the households. The households were then categorised into seven classes based on their total scores on the index. Excluding the top 10% of the population (0.5 points on the composite index, in a maximum total score of 5.0) the remaining index score was divided into six equal classes (of 0.75 points each), to avoid arbitrariness, giving us the following class composition in the population: i. ii. iii. iv. v. vi. vii. Lower lower class Middle lower class Upper lower class Lower middle class Middle middle class Upper middle class Upper class (LLC) (MLC) (ULC) (LMC) (MMC) (UMC) (UC) < 0.75 0.75- 1.5 1.5-2.25 2.25-3.0 3.0- 3.75 3.75- 4.5 > 4.5

Table 3 presents the weights assigned to each of the sub-index and the mean score achieved by the different classes on each index, and in total. The Table presents an interesting trend with the Lifestyle Index, based on a households possession of consumer durables carrying the maximum weight, followed by indices of income, education, housing and occupation. Is lifestyle the most distinctive factor in class differentiation? The answer would probably be in the affirmative if we look at the factors differentiating the UMC and UC, where the main difference between the two is for the Lifestyle Index. In this regard, it is also worth noting that housing and lifestyle indices are the ones that carry scores even for the lowest of classes, even when they score poorly on the income, occupation and education indices (see Table 3). Expectedly, as we go up the classes, households begin to score on all indices6. The upper middle class is almost
6

See Figure A2 in the Annexure for the graphic representation of the composite index, and its components, for each class. See also Figure A3 for thepercentage shareof each component index inaparticular classs total score on the composite index. Interesting to see in the latter graph is the appearance, and increasing size, of indices, as opposedtoafewindiceshavingamajorshare.

14

similar to the upper class, as can be seen from the similar mean scores on the indices for income, education and occupation. The differences in the lifestyle and housing indices, however, separate the top two classes, as can be seen in Table 3. Table3:EstimationofMiddleClassthroughaWeightedCompositeIndex
Income Education 1.005 0.000 0.014 0.140 0.528 0.564 1.004 1.005 0.159 Indices Housing Occupation WeightsappliedtoeachIndex1 0.955 0.865 Meantotalforeachindex1 0.267 0.000 0.462 0.336 0.526 0.483 0.496 0.720 0.661 0.683 0.682 0.862 0.870 0.865 0.417 0.300 Lifestyle 1.125 0.186 0.268 0.324 0.364 0.451 0.494 0.802 0.275 Total 5.000 0.453 1.097 1.802 2.662 3.301 4.092 4.592 1.361

Classes 1.050

Lowerlower Middlelower Upperlower Lowermiddle Middlemiddle Uppermiddle Upper Total

0.000 0.016 0.329 0.554 0.941 1.050 1.050 0.211

Source:CalculatedfromPSLM20072008.Note:1.Inordertoavoidtoomanydecimalsforeaseincomprehension,theweights were multiplied by 5 to form the composite index score. The weights used for the various indices as calculated by PCA are: Income 0.210; Education 0.201; Housing 0.191; Occupation 0.173; Lifestyle 0.225; and Total 1.00. The procedure does not in anywayaffecttheclassificationofthesample.

Class structure as calculated by the composite weighted index shows that a large majority of the people in Pakistan falls in the lower classes, be it lower lower class (LLC), middle lower class (MLC) or upper middle class (ULC), as can be seen from Table 4. This trend is mainly due to the rural areas that are predominantly concentrated in the lower classes. A moderate proportion (33 per cent)in the urban areas, however, can be categorised as belonging to the middle class (LMC, MMC and UMC put together). The biggest class, nevertheless, remains the LLC be it the urban or the rural areas (Table 4), and because of the very low index score it would not be inappropriate to label this category as vulnerable. If we look at the index scores of the MLC, which is the second largest class, in Table 3, we find the households to be scoring on all subindices, unlike the LLC. These households can be considered the aspirants for upward mobility. The next class, ULC, shows a marked improvement on all sub-indices (Table 3) and can rightfully be called a class of climbers. They are on the border of the middle class, aiming to get elevated. Even if we look at their total index score, they are found to have crossed the mean score of the sample (1.36) with a score of 1.8 (see Table 3), and with a little arbitrariness in demarcating the class boundaries can be a part of the middle class. 15

Table4:SizeofDifferentClassesthroughaWeightedCompositeIndexinPakistan
Class Lowerlower(LLC) Middlelower(MLC) Upperlower(ULC) Lowermiddle(LMC) Middlemiddle(MMC) Uppermiddle(UMC) Upper(UC) Total Categorisationfor middleclass1 Vulnerable Aspirants Climbers Fledglingmiddleclass Hardcoremiddleclass Elitemiddleclass Privileged Proportion(%) Total Urban 41.9 23.6 23.0 21.8 15.8 20.8 8.5 12.5 4.3 8.1 6.0 12.3 0.4 0.9 100.0 100.0 Rural 55.2 23.9 12.3 5.7 1.6 1.3 0.1 100.0 Numbers(inmillions)2 Total Urban Rural 83.7 20.1 63.6 41.9 16.4 25.6 28.5 15.9 12.6 16.3 9.4 6.8 6.7 5.2 1.5 9.5 8.0 1.5 0.6 0.6 0.0* 111.6 187.2 75.6

Source:CalculatedfromPSLM20072008 Note:1.CategoriesadaptedfromSridharan(2004)andTharoor(2010). 2.ForthemethodemployedtocalculatethesenumbersseeA5. 3.*signifiesvaluelessthan0.1.

The middle class, as observed earlier, seems to be more of an urban phenomenon (Table 4) which is not surprising in the light of the poor literacy levels and the large association with manual jobs in the rural areas. Even in cases where the households score well on the income and housing indices, their scores go down because they perform badly on the education and occupation indices. Among the middle classes the LMC, termed as the Fledgling Middle Class, constitutes the largest share (Table 4). From the urban areas, at least, it appears that once a household enters the middle class category it is more likely to graduate to the UMC, that is the Elite Middle Class, than to remain in the Hard-core Middle Class category as reflected by the higher share for the former than the latter. As expected, the size of the UC is small, being even smaller in the rural areas than the urban. Sridharan (2004) advocates the inclusion of ULC and UC while demarcating the boundaries for what he refers to as the Broadest Middle Class. Considering the stringent measurement method suggested to estimate the middle class in this paper, it makes sense to include at least the ULC in what is referred to as the Expanded Middle Class here (see Table 5). This becomes even more logical for the reason stated above regarding ULC having a composite index score crossing the midline. Using the Strict Middle Class category the middle class is found to be comprising one-fifth of the country, a proportion that increases to one-third if we take into account only the urban areas. Adding the Climbers to this estimate boosts the proportion of those in the middle class to 35 per cent, swelling the urban middle class size to over half the population (54 per cent), as can be seen from Table 5. The Broadest Middle Class shows only a 16

marginal increase in the share of the middle class in the total population as those in the Privileged class comprise a very small proportion (Table 4).

Table5:SizeoftheMiddleClassinPakistanUsingaWeightedCompositeIndex
StrictMiddleClass (LMC+MMC+UMC) ExpandedMiddleClass1 (LMC+MMC+UMC+ULC) BroadestMiddleClass1 (ULC+LMC+MMC+UMC+UC) Proportion(%) Total Urban Rural 18.8 32.9 8.6 34.6 53.7 20.9 35.0 54.6 21.0 Numbers(inmillions)2 Total Urban Rural 32.5 22.6 9.8 61.0 38.5 22.4 61.6 39.1 22.4

Source:CalculatedfromPSLM20072008 Note:1.AscategorisedbySridharan(2004). 2.ForthemethodemployedtocalculatethesenumbersseeA5.

Numbers are easier to comprehend than proportions. Hence, if we look at the numbers constituting the middle class in Pakistan we find the numbers to increase from 34 million (Strict Middle Class) to over 61 million (Broadest Middle Class) by widening its scope (Table 5).The middle class being more urban centric, as we saw in Table 4, amounts to nearly 39 million people if we go by the measure provided by the Broadest Middle Class definition in Table 5. The present paper, however, prefers to subscribe to the Expanded Middle Class estimate as representing the middle class in Pakistan. The UC has its peculiar characteristics and position in any society and should not be grouped with other classes. On the other hand, inclusion of the Climbers in the expanded middle class is rationalised on the basis of closeness between the two classes. It is from this class that the households make the transition from having a deficit budget to a surplus one, a characteristic deemed essential by many (including, Brandolini, 2010; Birdsall, Graham and Pettinato, 2000; Andersen, 1992) to be classified as the middle class7. The size of the middle class in Pakistan, using the Expanded Middle Class categorisation, comes up to a substantial 61 million people. It would be of interest to note how does the size of the middle class estimated through the composite index compare with the most common prevailing method of measuring it. This is based on the gradation of households per person per day expenditure (Definition Thirteen in
Forthegraphicrepresentationofhouseholdsincome,expenditureandthebalancebetweentwoseeA4.Ascan be seen from the graph, the Vulnerable and the Aspirants have a deficit budget and the trend reverses for the Climbers,justifyingtheinclusionofthelatterintheExpandedMiddleClass.
7

17

Table 2). As opposed to the latter, the composite index shows a much larger LLC and MLC, as can be seen in Figure 1. The only other class comparatively larger in the composite index is that of the UMC. The size of the UC, interestingly, remains the same in both the measures at a low 0.4 per cent. Despite giving an estimate for the middle class that is much bigger and hence more attractive, this paper argues that measuring the middle class only on the basis of expenditure or income is insufficient as it does not encompass all the factors that contribute towards making a household a middle class household. It is also premised that class is a much more permanent category to be affected by the mere fluctuation of income or expenditure. We would not go to the extreme of subscribing to the notion that class is permanent, as class mobility does and should take place, but it is not as transient to be altered by fickle changes in income or expenditure.
Figure1:ComparisonoftheSizeoftheClassesasEstimatedbytheWeighted CompositeIndexandthePrevailingMeasure1
(%)

UC

0.4

0.4

UMC

1.3

6.0

PrevailingMeasure CompositeWeightedIndex

MMC

9.8

4.4

LMC

32.4

8.5

ULC

34.8

15.8

LLC& MLC

21.3

64.9

55

35

15

25

45

65

Source:CalculatedfromPSLM20072008 Note:1.DefinitionThirteeninTable2. 2. Classes: Lower lower (LLC), Middle lower (MLC), Upper lower (ULC), Lower middle (LMC), Middle middle(MMC),Uppermiddle(UMC),Upper(UC)

Can we consider a household with only manual workers or with no individual having college education but having over PPP $2 per person per day income middle class in todays world? Following how the middle class is generally perceived, the answer in all probability would be in the negative. Figure 2 presents the comparison between the two measures, the composite index and the prevailing definition, on the nature of occupation and presence of a college graduate in 18

the households belonging to different classes. The estimate through the weighted composite index appears more appropriate with those in the elite middle class and the privileged class being employed solely in non-manual occupations and having at least one college graduate in the household (Figure 2). On the contrary, the prevailing definition shows a substantial proportion doing manual jobs and having no college graduate in the household, something that cannot be associated with the middle class nowadays. Such a trend is, however, expected when only income or expenditure form the basis of the classification.

Figure2:NatureofOccupationandLevelofEducationbyClassesthroughtheWeighted CompositeIndexandthePrevailingDefinition2

A:PresenceofaCollegeGraduateintheHousehold
100 90 80 50.8 70 68.3 69.0 60 50 40 30 49.2 20 31.7 31.0 10 0 LC LMC MMC UMC UC 0.1 LLC MLC ULC LMC MMC UMC UC 47.4 43.9 94.7 79.3 Has 100.0 98.6 Graduate No Graduate 86.1 99.9 100.0 52.6 56.1 5.3 20.7 1.4 13.9

B:NatureofOccupation
100 90 80 70 60 50 40 30 20 10 0 LC LMC MMC UMC UC 77.1 55.6 36.9 25.4 24.1 22.9 44.4 63.1 74.6 75.9 Non 100.0 manual Manual 61.2 44.2 16.8 21.0 0.3 LLC MLC ULC LMC MMC UMC UC 38.8 55.8 83.2 79.0 99.7 100.0

PrevailingDefinition

WeightedComposite Index

PrevailingDefinition

WeightedCompositeIndex

Source:CalculatedfromPSLM20072008 Note:1.Classes:Lowerlower(LLC),Middlelower(MLC),Upperlower(ULC),Lowermiddle(LMC),Middlemiddle(MMC),Uppermiddle(UMC), Upper(UC).2.DefinitionThirteeninTable2.

For the sake of comparison, it would have been an interesting exercise to apply the proposed methodology to estimate the size of the middle class to other countries in the region as well. Lack of access to household level data in these countries, however, hampered this effort at present. Nevertheless, if we compare the size of the middle class in the South Asian region estimated by various current studies using the prevailing definition based on expenditure levels (Definition Thirteen in Table 1) we find the size of the middle class to be the largest in Pakistan with the exception of Sri Lanka (see A-9). In a comparable year of 2004-05 the size of the middle class in Pakistan was 40 per cent of the total population while that of India was 25 per cent (see A-9).

19

Where is the Middle Class Present in Pakistan? : Region and Occupation

Once the middle class has been identified, it is of interest to see where those belonging to this class are present, geographically and sectorally. The provinces of Punjab and Sindh, having over 36 per cent middle class households, fare better than the provinces of Khyber Pakhtunkhwa (KP) and Balochistan that marginally lag behind at 32 and 28 per cent, respectively (see Table 6). When we look at the regional trends, we observe the national level pattern to be replicated at the provincial level as well. The size of the middle class is estimated to be much bigger in the urban areas as compared to the rural areas in all four provinces (Table 6), strengthening the argument that the middle class is more of an urban phenomenon. If we look at the inter-provincial differences, we find the size of the middle class to be positively associated with the proportion of the urban population in that province. The province of Sindh has the highest proportion of urban population and that of the middle class (56 per cent) in the urban areas as well, as can be seen in Table 6. The provinces of Punjab, Balochistan and KPK have progressively smaller share of the population living in the urban areas and their size of the middle class reflects it through the estimated measures for the class at 55 per cent, 50 per cent and 49 per cent, respectively. Increasing urban concentration seems to aid the increase in the size of the middle class. On the contrary, the size of the middle class in the rural areas in all four provinces, as shown in Table 6, is much smaller than their urban counterparts. Table6:RegionalandProvincialClassesEstimatedbyWeightedCompositeIndex (%)
Urban Urban Urban Urban Class Total Punjab Rural Total Sindh Rural KhyberPakhtunkhwa Total Rural Total Balochistan Rural

Lowerlower Middlelower Upperlower Lowermiddle Middlemiddle Uppermiddle Upper Total

38.6 21.2 51.6 45.0 24.7 64.3 23.7 22.4 24.7 18.4 18.4 18.2 17.1 20.6 14.5 14.2 20.6 8.1 8.4 12.6 5.3 9.7 12.3 7.2 4.9 8.9 1.8 5.3 9.3 1.4 6.5 12.8 1.9 7.2 14.1 .6 0.7 1.4 0.1 0.3 0.6 0.1 100.0 100.0 100.0 100.0 100.0 100.0 SizeofMiddleClass1 36.9 54.9 23.5 36.4 56.3 17.1

41.4 26.5 15.3 8.6 3.0 4.9 0.2 100.0 31.8

26.4 24.3 20.4 12.2 5.5 10.6 0.6 100.0 48.7

50.3 27.9 12.2 6.5 1.4 1.6 0.1 100.0

47.8 26.1 59.9 23.3 23.2 23.4 15.3 22.1 11.5 6.8 12.6 3.6 3.0 6.2 1.2 3.6 9.4 0.4 0.1 0.4 0.0 100.0 100.0 100.0 (%) 21.7 28.7 50.3 16.7

Source:CalculatedfromPSLM20072008 Note:BasedontheExpandedMiddleClass.

20

Discussion on class is incomplete without investigating the relationship between class and occupation, as also stressed by Banerjee and Duflo(2007), Goldthorpe and McKnight (2006), Wright (1997), Erikson and Goldthorpe (1992). If we look at the nature of occupations within each of the estimated classes, we see that the LLC comprises mostly occupations like agriculture and fishery workers, crafts and related trade workers, plant and machine operators and assemblers and the elementary occupations, as can be seen in Table 7. From the ULC onwards the occupational distribution exhibits a larger spread that tapers again for the highest two classes, the UMC and UC, which are confined to occupation like: armed forces; legislators, senior officials and managers; professionals; technicians and associate professionals; clerks; and service workers and sales workers (Table 7). Thus, as we go up the class ladder the occupational share shifts from manual to non-manual jobs, which to some is the very essence of the middle class8. Table7:OccupationalDistributionwithinClassesbyWeightedCompositeIndex (%)
Occupation Armedforces Legislators,seniorofficialsandmanagers Professionals Techniciansandassociateprofessionals Clerks Serviceworkers&shop/marketsalesworkers Agricultureandfisheryworkers Craftsandrelatedtradeworkers Plantandmachineoperatorsandassemblers Elementaryoccupations Total
Source:CalculatedfromPSLM20072008.

ClassesthroughWeightedIndex Middle Lower Middle Middle Lower Middle Upper Middle Upper Lower Upper 3.2 30.6 37.1 9.7 1.6 17.7 0.0 0.0 0.0 0.0 100.0 Lower Lower

0.0 0.0 0.0 0.0 0.0 0.0 33.2 9.0 11.9 45.9 100.0

0.8 0.3 2.9 7.0 3.6 31.0 17.0 8.5 9.2 19.6 100.0

0.7 0.9 4.6 10.0 7.6 34.1 14.9 8.5 7.1 11.7 100.0

1.1 2.9 18.3 19.3 14.5 30.0 5.9 2.6 2.1 3.3 100.0

0.6 8.2 15.7 14.2 11.8 28.9 8.6 5.0 3.1 3.9 100.0

1.3 20.7 26.4 21.3 12.9 17.1 0.1 0.0 0.0 0.0 100.0

Worth noting in Table 7 are the occupational distributions for the elite middle class and the privileged class that are heavily tilted towards professionals and those associated with services. This trend, observed for the middle class globally, can be best summarised in the words of Banerjee and Duflo (2007: 21) when they say,
To find how class distribution varies by different occupations see table A6 in the Annexure. The share of the middle class, specifically the elite middle class, is much higher among professionals, legislators, senior officials and managers. Elementary occupations and those related to agriculture and crafts are dominated by the lower classes.TheUCisalmostconfinedtothearmedforces,professionals,legislators,seniorofficialsandmanagers.
8

21

Nothing seems more middle class than the fact of having a steady well-paying job. While there are many petty entrepreneurs among the middle class, most of them do not seem to be capitalists in waiting.. If they could only find the right salaried job, they might be quite content to shut their business down..Perhaps the sense of control over the future that one gets from knowing that there will be an income coming in every month -- and not just the income itselfis what allows the middle class to focus on building their own careers and those of their children. Is there any industrial variation vis--vis distribution of middle class in Pakistan? Table 8 shows that two industries, namely, of wholesale, retail trade, hotel and restaurants, and those comprised of community, social and personal services, have a heavy share of the expanded middle class. It is interesting to find out that a noticeable share of the elite middle class and the privileged class is employed in manufacturing, financing, insurance, real estate and business services (Table 8). Not surprisingly, the LLC as a big share employed in agriculture, forestry, hunting and fishing9. Table8:IndustrialDistributionwithinClasses (%)
ClassesthroughWeightedIndex Middle Lower Middle Middle Lower Middle Upper Middle Upper Lower Sectors Upper 1.6 1.6 12.9 1.6 1.6 22.6 1.6 14.5 41.9 100.0 Lower Lower

Agriculture,forestry,huntingandfishing Miningandquarrying Manufacturing Electricity,gasandwater Construction Wholesaleandretailtrade,&hotel/restaurants Transport,storageandcommunication Financing,insurance,realestate&businessservices Community,socialandpersonalservices Total
Source:CalculatedfromPSLM20072008.

42.9 0.8 9.3 0.6 19.1 4.5 12.0 0.2 10.5 100.0

19.4 0.5 8.8 1.5 8.3 24.5 9.1 0.8 27.1 100.0

15.5 0.0 9.3 2.4 5.1 29.4 6.9 2.1 29.1 100.0

6.2 0.8 7.6 2.6 2.5 25.6 4.3 3.6 46.8 100.0

8.9 0.5 9.7 2.6 2.7 27.0 4.8 6.9 36.8 100.0

1.4 0.9 11.6 2.7 2.1 15.5 4.9 9.4 51.5 100.0

Relationship with the means of production has been a recurrent theme in the literature on class. As discussed earlier, the Marxian tradition considers this as the very foundation of class
TofindoutthedistributionofclassesineachindustryseefigureA7intheAnnexure.Withabigproportionofthe population falling in the LLC it is expected to find them having big shares in almost all of the industries. Worth noticing in this table, however, is the dominance of the expanded middle class in the industry comprising financing, insurance, real estate and business services, proving the observation of Banerjee and Duflo (2007), statedabove,validinPakistanaswell.
9

22

formation. Table 9 aims at looking into this very relationship and presents the shares of different statuses in employment within different classes. Large shares of paid employees in the expanded middle class and the privileged class tend to negate the relationship postulated by Marx regarding ownership of the means of production and class. As can be seen from Table 9, paid employees constitute 78 percent of the households in the UC. The Marxian notion can be seen to play some role in the increasing share of those who are self-employed in the nonagricultural sector in the expanded middle class and those who employ more than 10 employees in the hard-core middle class, the elite middle class and the privileged class10. Table9:StatusofOccupationwithinClasses (%)
ClassesestimatedbycompositeweightedIndex Middle Lower Middle Middle Lower Middle Upper Middle Upper Lower Statusofoccupation Upper 0.0 3.2 22.6 74.2 0.0 0.0 0.0 0.0 0.0 100.0 Lower Lower

Employerwith<10employees Employerwith>10employees Selfemployednonagriculture Paidemployee Unpaidfamilyworker Owncultivator Sharecropper Contractcultivator OwnLivestock Total
Source:CalculatedfromPSLM20072008.

0.2 0.1 8.7 59.8 0.1 18.1 6.7 2.7 3.7 100.0

0.4 0.2 25.8 57.6 0.1 11.9 1.3 0.8 1.9 100.0

0.7 0.2 28.8 56.0 0.0 12.0 0.4 1.0 0.9 100.0

1.1 1.4 27.5 64.5 0.0 4.7 0.3 0.3 0.3 100.0

1.6 3.9 27.0 60.1 0.0 6.3 0.0 0.6 0.5 100.0

1.0 4.0 16.9 78.0 0.0 0.1 0.0 0.0 0.0 100.0

Conclusions Social realities are difficult to define, even tougher to measure and quantify, and the term middle class is no exception. Just like poverty, it is variously defined in different countries at different levels of development. The definitions differ with the variance in approach and rationale applied in addressing the concept, thus leading to a lack of consensus in what the term actually means. The only agreement regarding the term middle class is that it is the class between the lower and the upper class but where are these lines demarcated among them remain a debatable issue. Classifications made on cut-off lines based on income or expenditure
10

For distribution of classes by status of occupation, see figure A8. It would not be wrong to infer from the table thattheexpandedmiddleclassesandtheprivilegedclassaremainlyfoundamong:paidemployees;selfemployed in the nonagricultural sector; and employer with more than ten employees. The last two statuses in this regard justifytheMarxianstanceregardingclassandrelationshipwiththemeansofproduction.

23

continue to be the most commonly used way of measuring classes, and arbitrariness remains the hallmark of all these definitions. This arbitrariness is reflected in the range of estimates given for the middle class, varying from zero per cent to 60 per cent. The present paper, however, considers these definitions inadequate to capture the whole concept of the middle class, and suggests a weighted composite index to estimate its size.

The suggested measure is a composite of five weighted sub-indices of factors believed to be important for being part of the middle class, namely, education, occupation, income, lifestyle and housing. Using the expanded middle class concept, Pakistan is estimated to have a middle class that is around 35 per cent of the total population, which approximates to a substantial 61 million. The middle class is found to be more of an urban phenomenon with its size being much larger in the urban areas at both the national and the provincial levels. A striking feature, however, is the association found between the professional occupations and the upper middle and upper classes. This fits in well with the general belief that professional occupations constitute, what in this paper has been referred to as, the elite middle class. Other white-collar occupations are taken up by the hard-core and fledgling middle class, and the manual occupations comprise the vulnerable and the aspirants. Likewise, the middle class is expected to have sufficient resources to fulfil all their needs and at least some of their wants, and have a surplus to save for themselves. This criterion works for the estimated middle class in this paper as well and provides the very rationale of including climbers, who show a surplus/saving trend, in the expanded middle class category.

It would be of interest to carry out a comparative study in the South Asian region using the proposed multidimensional approach to gauge the actual size of the middle class. However, using the existing definitions, when compared to its neighbours, baring Sri Lanka, Pakistan has a bigger middle class than all other countries including India. Of course, the Indian middle class would be bigger in numbers given its much larger population size but it comprises a smaller proportion that falls in the middle class category. Hence, harnessing the gains that are associated with having a big middle class are potentially available to the country. Is the middle class shrinking in Pakistan? Due to the recent inflationary trends, it is a fear that is much voiced in Pakistan, as in many other countries. For the period covered, however, applying the most

24

commonly used existing definitions, the answer to this question is in the negative. The middle class in Pakistan has actually grown over time (see A-9). Likewise, the multidimensional definition of class proposed in this paper has a sense of stability linked to it, making the middle class less susceptible to fluctuations in income or expenditure. By differentiating between the concept of middle class from that of middle income, we can understand why the inflationary trends do not have an immediate effect on class structure, and hence the middle class. Otherwise as well, since all the factors linked to being a part of the middle class, reflected in their inclusion in the proposed weighted composite index, show an increasing trend over time in Pakistan, the size of the middle class is bound to increase in the country. The scenario cannot be better described than in the words of Burke (2010) when he uses the analogy of the car Suzuki Mehran for the middle class in Pakistan and says, In Pakistan, the hierarchy on the roads reflects that of society. If you are poor, you use the overcrowded buses or a bicycle. Small shopkeepers, rural teachers and better-off farmers are likely to have a $1,500 Chinese or Japanese motorbike. Then come the Mehran drivers. A rank above them, in air-conditioned Toyota Corolla saloons, are the small businessmen, smaller landlords, more senior army officers and bureaucrats. Finally, there are the luxury four-wheel drives of feudal landlords, big businessmen, expats, drug dealers, generals, ministers and elite bureaucrats. The latter may be superior in status, power and wealth, but it is the Mehrans which, by dint of numbers, dominate the roads.

25

References Acemoglu, D., and Robinson., J. (2003). Political Origins of Dictatorship and Democracy. Massachusetts: MIT. Acs, G., and Loprest, P. (2005). Who Are Low-Income Working Families? Washington DC: Urban Institute. ADB.(2010). The Rise of Asias Middle Class. Manila: Asian Development Bank. Aho, K. (2009). Who in the World is Middle Class? New York: Money Matters. Andersen, G. E. (1992). Post-Industrial Class Structures: An Analytical Framework: Center for Advanced Study in the Social Sciences, Institute in Madrid: Working Paper 1992/38. Atkinson, A., and Bourguignon, F. (1982). The Comparison of Multidimensioned Distributions of Economic Status. Review of Economic Studies, 49, 183-201. Banerjee, A., and Duflo, E. (2007). What is Middle Class About the Middle Classes Around the World? Massachusetts: MIT. Bardhan, P. (1994). The Political Economy of Development in India. New Delhi: Oxford University Press. Bardhan, P. (1989). The Third Dominant Class. Economic and Political Weekly, 155-156. Basu, I. (2003). India's Growing Urge to Splurge. Asia Times Online. Retrieved December 12, 2010, from the World Wide Web: http://www.atimes.com/atimes/South_Asia/EH22Df01.html Beteille, A. (2001). The Social Character of the Indian Middle Class. In I. Ahmad & H. Reifeld (Eds.), Middle Class Values in India and Western Europe. New Delhi: Konrad Adenauer Foundation, Social Science Press. Bhandari, L. (2010). Neither Middling Nor Muddled: A Study of the Indian Middle Classes. Manila: Paper presented during the Workshop on Asias Middle Class held at ADB Headquarters, on 2728 May. Birdsall, N. (2010). The (Indispensable) Middle Class in Developing Countries; or, The Rich and the Rest, Not the Poor and the Rest: Center for Global Development: Working Paper 207. Birdsall, N., Graham, C., and Pettinato, S. (2000). Stuck In The Tunnel: Is Globalization Muddling The Middle Class? Center on Social and Economic Dynamics. Bourguignon, F., and Chakravarty, S. (2003). The Measurement of Multidimensional Poverty. Journal of Economic Inequality, 1, 25-40.

26

Brandolini, A. (2010). On the Identification of the Middle Class: Bank of Italy, Department for Structural Economic Analysis. Brandolini, A., Magri, S., and Smeeding, T. (2010) Asset-Based Measurement of Poverty. Journal of Policy Analysis and Management 29(2), 267-284. Brzezinski, M. (2010).Income Affluence in Poland. Social Indicators Research, 99(2), 285-299. Burke, J. (2010). Letter from Karachi. Prospect(169). Chun, N. (2010). Middle Class Size in the Past, Present, and Future: A Descriptive Analysis of Distributional Trends and Projections: Asian Development Bank: ERD Working Paper. Cole, G. D. (1950). The Concept of the Middle Class. The British Journal of Sociology, 1(4), 275-290. Collier, P. (1999). Ethnicity, Politics, and Economic Performance: Mimeo. Dahrendorf, R. (1959). Class and Class Conflict in Industrial Society. Stanford: Stanford University Press. Easterly, W. (2001).The Middle Class Consensus and Economic Development. Journal of Economic Growth, 6, 317-335. Eisenhauer, J. G. (2008). An Economic Definition of the Middle Class. Forum for Social Economics, 37(2), 103-113. Erikson, R., and Goldthorpe, J. (1992). The Constant Flux. Oxford: Clarendon Press. Esteban, J., and Ray, D. (1999). Conflict and Distribution. Journal of Economic Theory, 87, 379-415. Giddens, A. (1973). The Class Structure of Advanced Societies. London: Hutchinson. Gigliarano, C., and Mosler, K. (2009). Measuring Middle-Class Decline in One and Many Attributes: Universita Politechnica Delle Marche, Quaderno Di Ricerca n. 333. Gilbert, D. (2003). The American Class Structure. United States: Thomson-Wadsworth. Goldthorpe, J. (1987). Social Mobility and Class Structure in Modern Britain. Oxford: Clarendon Press. Goldthorpe, J., and Marshall, G. (1992). The Promising Future of Class Analysis: A Response to Recent Critiques. Sociology, 26(3), 381-400. Goldthorpe, J., and McKnight, A. (2006). The Economic Basis of Social Class. In S. Morgan & D. B. Grusky & G. S. Fields (Eds.), Mobility and Inequality: Frontiers of Research from Sociology and Economics. California: Stanford University Press.

27

Goldthrope, J. (2010). Analysing Social Inequality: A Critique of Two Recent Contributions from Economics and Epidemiology. European Sociological Review, 26(6), 731-744. Grant, G. (1983). The State and the Formation of a Middle Class: A Chilean Example. Latin American Perspectives, 2(3), 151-170. Horrigan, M., and Haugen, S. (1988). The Declining Middle-Class Thesis: A Sensitivity Analysis. Monthly Labor Review, 111(5), 3-13. Kolm, S. C. (1977). Multidimensional Egalitarianisms. Quarterly Journal of Economics, 91, 1-13. Landes, D. (1998). The Wealth and Poverty of Nations. New York: Norton. Lenski, G. (1966). Power and Privilege. New York: McGraw-Hill. Leventouglu, B. (2003). Social Mobility, Middle Class and Political Transitions. Stony Brook: Department of Political Science, Stony Brook University. McKinsey, T. (2010).Capturing the Worlds Emerging Middle class. New York: McKinsey Quarterly. Milanovic, B., and Yitzhaki, S. (2002). Decomposing World Income Distribution: Does the world have a middle class? Review of Income and Wealth, 48(2), 155-178. Moore, B. (1966). Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World. Boston: Beacon Press. Mosca, G. (1939). The Ruling Class. New York: McGraw-Hill. Nehru, V. (2010).Is there a Middle Class in Asia? World Bank: Poverty Reduction and Economic Management and Private and Financial Sector Department for the East Asia and Pacific Region. Olin, E. Wright, E.O. (1979), Class Structure and Income Determination. New York: Academic Press. Ossowski, S. (1963).Class Structure in the Social Consciousness. London: Routledge & Kegan Paul. Pareto, V. (1963).The Mind and Society (Vol. 1-2). New York: Dover. Peichl, A., Schaefer, T., and Scheicher, C. (2008). Measuring Richness and Poverty: A Micro Data Application to Europe and Germany: IZA Discussion Paper no. 3790. Ravallion, M. (2010).The Developing Worlds Bulging (but Vulnerable) Middle Class. World Development, 38(4), 445-454. Schumpeter, J. (1951). Imperialism and Social Classes. New York: Augustus M. Kelley. Singh, N. (2005). The Idea of South Asia and the Role of the Middle Class. Santa Cruz Center for 28

International Economics, University of California, Santa Cruz: Paper 05-08. Sorenson, A. (2000). Toward a Sounder Basis for Class Analysis. The American Journal of Sociology, 105(6), 1523-1558. Sridharan, E. (2004). The Growth and Sectoral Composition of India's Middle Class: Its Impact on the Politics of Economic Liberalization. India Review, 3(4), 405-428. Stearns, P. N (1979).The Middle Class: Toward a Precise Definition. Comparative Studies in Society and History, 21(3), 377-396. Tharoor, S. (2005). Who is this Middle Class? The Hindu Online, accessed on January, 20, 2011 at
http://tharoor.in/articles/whoisthismiddleclass/

Thurow, L. (1987). A Surge in Inequality. Scientific American, 256, 30-37. Tilkidjiev, N. (1998). The Middle Class: The New Convergence Paradigm: Institute of Sociology, Bulgarian Academy of Sciences. UN (2009) World Population Prospects, Revision 2009. United Nations Population Division. USCB (2011).International Database. United States Census Bureau. Warner, L. W. (1960). Social Class in America. New York: Harper & Row. Wheary, J. (2005). Measuring the Middle Assessing What It Takes to Be Middle Class. New York: Demos: Working Paper. Wright, E. O. (2003). Social Class. In G. Ritzer (Ed.), Encyclopedia of Social Theory: Sage Publication. Wright, E. O. (1997). Class Counts: Comparative Studies in Class Analysis. Cambridge: Cambridge University Press. Wright, E. O. (1985). Classes. London: Verso Press. Wright, E. O. (1979). Class Structure and Income Determination. New York: Academic Press. Yuan, Z., Wan, G., and Khor, N. (2011). The Rise of the Middle Class in the Peoples Republic of China. Manila: ADB Economics: Working Paper 247.

29

Annexure
A1:ClassStructureBasedonIncomeandOccupation

Source:Gilbert(2003:8)

A2:ContributionofEachSubindexintheTotalWeightedComposite IndexforEachClass
5.0 4.5 4.0 3.5 3.0 2.5 2.0 0.140 1.5 1.0 0.5 0.0 0.267 0.186 LLC 0.014 0.336 0.016 0.462 0.268 MLC 0.483 0.329 0.526 0.324 ULC 0.554 0.496 0.364 LMC 0.661 0.451 MMC 0.682 0.494 UMC 0.802 UC 0.720 0.941 1.050 0.870 Education Occupation Income House Lifestyle 0.528 0.564 0.862 0.683 1.050 1.004 0.865 1.005

Source:CalculatedfromPSLM20072008.

30

A3:ShareoftheSubindicesintheTotalScoreforEachClassEstimated ByWeightedCompositeIndex (%)

100 90 80 70 60 50 42.14 40 30 20 10 0 LLC Lifestyle MLC House ULC LMC Income MMC Occupation UMC UC Education 41.11 24.46 17.99 29.21 18.62 20.03 18.95 16.66 12.08 17.46 20.80 58.89 30.62 26.79 20.70 1.50 18.25 28.52 25.66 22.86 27.05 21.07 18.84 1.27 7.76 19.85 17.08 24.53 21.88

13.67

13.67

Source:CalculatedfromPSLM20072008.

A4:MeanAnnualHouseholdIncomeExpenditureBalancebyDifferent ClassesasEstimatedbyWeightedCompositeIndex (in000Pak.Rs.)
200 UC,184

150

UMC,108 100 MMC,69

50 LMC,22 ULC,13 0 LLC,6 MLC,19 50

Source:CalculatedfromPSLM20072008.

31

A5:MethodtoCalculateNumbersinEachClassfromtheProportionsAchievedfrom WeightedCompositeIndex
The middle class status is assigned to a household and notan individual.Hence, theproportion estimated to be in each class is in fact proportion of households and not the number of persons in a population belonging to any particular class. Instead of a crude estimation of numbers from the calculated proportions in all classes, size of every class was measured according to the respective mean household size found in each class. The midyear populationof187.2millionin2011wasusedasthebaseyearforestimatingthesizeofthemiddleclass.Therural urban ratio found in the sample, andapplied in this calculation, was 60:40. The size of the middleclass in Pakistan hasgenerallyshownagrowingtrend(seeA9)soapplyingthe20072008estimatesto2011populationcangivean underestimation/minimum size but not an overestimation. Otherwise too, the composite index would be only marginallyaffectedbyshortrunchangesinincomeorexpenditure.

Themeanhouseholdsizeforeachclassusedtocalculatethenumbersindifferentclasses,byregion,isasfollows: MeanHouseholdSize Class Total Urban Rural Lowerlower(LLC) 7.43 7.53 7.40 Middlelower(MLC) 6.80 6.66 6.90 Upperlower(ULC) 6.69 6.75 6.62 Lowermiddle(LMC) 7.08 6.66 7.76 Middlemiddle(MMC) 5.72 5.65 5.99 Uppermiddle(UMC) 5.94 5.75 7.19 Upper(UC) 5.60 5.49 6.80 Total 6.97 6.68 7.18
Source:CalculatedfromPSLM20072008.

A6ClassDistributionwithinOccupations(%)
ClassesestimatedbycompositeweightedIndex Middle Lower Lower Middle Middle Middle Upper Middle Upper Lower Upper Lower Lower Occupations Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Armedforces Legislators,seniorofficialsandmanagers Professionals Techniciansandassociateprofessionals Clerks Serviceworkers&shop/marketsalesworkers Skilledagricultureandfisheryworkers Craftsandrelatedtradeworkers Plantandmachineoperatorsandassemblers Elementaryoccupations
Source:CalculatedfromPSLM20072008.

0.0 0.0 0.0 0.0 0.0 0.0 67.0 51.2 59.5 75.0

34.8 2.6 11.1 21.4 16.5 37.8 17.0 24.0 22.8 15.8

21.7 6.4 13.7 24.1 26.7 32.5 11.7 18.7 13.7 7.4

18.8 11.3 29.2 24.8 27.5 15.4 2.5 3.0 2.2 1.1

5.8 16.4 12.9 9.5 12.3 7.6 1.9 3.0 1.6 0.0

15.6 57.2 30.1 19.7 17.6 6.3 0.0 0.0 0.2 0.0

2.9 6.1 3.1 0.6 0.2 0.5 0.0 0.0 0.0 0.0

32

A7:ClassDistributionwithinIndustries (%)
ClassesestimatedbycompositeweightedIndex Middle Lower Middle Middle Lower Middle Upper Middle Upper Lower Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Total Occupations Upper 0.0 1.1 0.6 0.5 0.1 0.6 0.1 3.4 0.8 Lower Lower

Agriculture,forestry,huntingandfishing Miningandquarrying Manufacturing Electricity,gasandwater Construction Wholesaleandretailtrade,&hotel/restaurants Transport,storageandcommunication Financing,insurance,realestate&businessservices Community,socialandpersonalservices

Source:CalculatedfromPSLM20072008.

70.4 56.3 42.7 17.6 72.5 11.7 56.1 5.4 18.5

15.7 17.2 20.1 21.1 15.5 31.2 21.1 8.8 23.5

9.9 1.1 16.5 26.5 7.6 29.4 12.6 18.4 19.7

2.1 11.5 7.3 15.2 2.0 13.7 4.2 16.5 17.0

1.6 3.4 4.8 7.8 1.1 7.5 2.4 16.5 6.9

0.3 9.2 8.0 11.3 1.2 5.9 3.4 31.0 13.5

A8:ClassesDistributionbyStatusofOccupation(%)
ClassesestimatedbyWeightedCompositeIndex Middle Lower Lower Middle Middle Middle Upper Middle Upper Lower Statusofoccupation Upper 0.0 2.1 0.6 0.6 0.0 0.0 0.0 0.0 0.0 Lower Lower

Employerwith<10employees Employerwith>10employees Selfemployednonagriculture Paidemployee Unpaidfamilyworker Owncultivator Sharecropper Contractcultivator OwnLivestock

13.6 8.4 19.8 42.0 57.1 59.7 88.6 74.3 72.4

15.2 6.3 29.0 20.0 42.9 19.5 8.4 10.7 18.7

22.7 4.2 25.3 15.2 0.0 15.3 2.3 11.2 6.8

19.7 17.9 13.0 9.4 0.0 3.2 0.7 1.9 1.0

15.2 25.3 6.6 4.5 0.0 2.2 0.0 1.9 1.0

13.6 35.8 5.7 8.2 0.0 0.1 0.0 0.0 0.0

Source:CalculatedfromPSLM20072008

33

A9:SizeandCompositionoftheStrictMiddleClassinthe SouthAsianRegion(%)
Pakistan2005 32.9 6.6 0.6

Pakistan2008

32.4

9.8

1.3

$2$4 Bangladesh 16.4 3.5 0.4 $4$10 $10$20

Nepal

16.7

5.3 0.9

SriLanka

37.8

18.7

2.7

India 0

20.5 10

4.15 0.4 20 30 40 50 60

Source:Chun(2010)exceptforPakistan2008,whichwascalculatedfromPSLM20072008.

34